---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
      name:  <unnamed>
       log:  C:\fsu2010\formateur\ajps\submission2\replication\analysis\stata_mxl_robustness.log
  log type:  text
 opened on:  21 Mar 2011, 13:48:47

. #delimit ;
delimiter now ;
. set mem 30m;

Current memory allocation

                    current                                 memory usage
    settable          value     description                 (1M = 1024k)
    --------------------------------------------------------------------
    set maxvar         5000     max. variables allowed           1.947M
    set memory           30M    max. data space                 30.000M
    set matsize         400     max. RHS vars in models          1.254M
                                                            -----------
                                                                33.201M

. clear matrix;

. clear mata;

. set more off;

. *     ***************************************************************** *;
. *     ***************************************************************** *;
. *       File-Name:  stata_mxl_robustness.do                             *;
. *       Date:       November 11, 2010                                   *;
. *       Author:     GG/MG                                               *;
. *       Purpose:    Robustness tests for AJPS paper                     *;
. *       Input File:     europe_all.dta, westeurope2_CMP.dta             *;
. *       Output File:    stata_mxl_robustness.log                        *;
. *       Data Output:    none                                            *;
. *       Previous file:  europe_all.dta, westeurope2_CMP.dta             *;
. *       Machine:        Matt's desktop                                  *;
. *     ****************************************************************  *;
. *     ****************************************************************  *;
. *     ****************************************************************  *;
. *       What if we dropped cases in Eastern Europe where the party of   *;
. *       the incumbent PM is not in the new legislature?                 *;
. *     ****************************************************************  *;
. use "C:\fsu2010\formateur\ajps\submission2\replication\analysis\europe.dta", clear;

. drop if cabinetcode == 2107;
(7 observations deleted)

. drop if cabinetcode == 2307;
(6 observations deleted)

. drop if cabinetcode == 2311;
(5 observations deleted)

. drop if cabinetcode == 2509;
(8 observations deleted)

. drop if cabinetcode == 2707;
(7 observations deleted)

. mixlogit choice largestparty_west party_seatshare_west medianparty1_west presidentparty_west previousPM_west previousPM_conflict_west previousPM_gain1_west previous_cabinet_gain1_west,  
>     rand (largestparty party_seatshare medianparty1 presidentparty previousPM president_investiture president_direct previousPM_conflict previousPM_gain1 previous_cabinet_gain1) 
>     ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -250.44035  (not concave)
Iteration 1:   log likelihood = -223.90829  
Iteration 2:   log likelihood = -221.74861  (not concave)
Iteration 3:   log likelihood =  -218.3123  (not concave)
Iteration 4:   log likelihood =  -217.9975  
Iteration 5:   log likelihood = -214.97803  
Iteration 6:   log likelihood = -214.30131  
Iteration 7:   log likelihood = -214.21904  
Iteration 8:   log likelihood =  -214.2168  
Iteration 9:   log likelihood = -214.21678  

Mixed logit model                                 Number of obs   =       2006
                                                  LR chi2(10)     =      20.10
Log likelihood = -214.21678                       Prob > chi2     =     0.0283

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~est |  -.7203934    2.44281    -0.29   0.768    -5.508212    4.067425
party_se~est |  -.1783579   .1295107    -1.38   0.168    -.4321943    .0754784
medianpa~est |   2.612985   1.596616     1.64   0.102    -.5163241    5.742294
presi~y_west |   2.628493   2.703181     0.97   0.331    -2.669645    7.926631
previ~M_west |   3.355454   2.886349     1.16   0.245    -2.301685    9.012594
previ~t_west |  -1.287004   3.522126    -0.37   0.715    -8.190243    5.616236
previousPM.. |    .084627   .7019521     0.12   0.904    -1.291174    1.460428
previous_c.. |   .0513543   .2644184     0.19   0.846    -.4668961    .5696048
largestparty |     3.1683   3.013599     1.05   0.293    -2.738245    9.074845
party_seat~e |     .31692   .1284644     2.47   0.014     .0651345    .5687055
medianparty1 |   -1.17821   1.531701    -0.77   0.442    -4.180287    1.823868
presidentp~y |   2.837327   3.555822     0.80   0.425    -4.131955    9.806609
  previousPM |  -.1734207   2.450663    -0.07   0.944    -4.976631     4.62979
president_~e |  -3.020047   2.284138    -1.32   0.186    -7.496874     1.45678
president~ct |   -4.05641     2.2991    -1.76   0.078    -8.562564    .4497431
previousP~ct |  -1.500828   3.279055    -0.46   0.647    -7.927657    4.926001
previousPM~1 |   .5624407    .700339     0.80   0.422    -.8101986     1.93508
previous_c~1 |  -.0158251    .260738    -0.06   0.952    -.5268622     .495212
-------------+----------------------------------------------------------------
SD           |
largestparty |  -5.239417   2.372228    -2.21   0.027    -9.888898   -.5899366
party_seat~e |  -.0110375   .0390879    -0.28   0.778    -.0876485    .0655734
medianparty1 |    .190017   .8616238     0.22   0.825    -1.498734    1.878769
presidentp~y |  -.1309576   1.086662    -0.12   0.904    -2.260775     1.99886
  previousPM |   1.486184   1.275835     1.16   0.244    -1.014406    3.986774
president_~e |   .0079059   1.528937     0.01   0.996    -2.988755    3.004567
president~ct |   -.122623   1.662659    -0.07   0.941    -3.381374    3.136128
previousP~ct |   2.074131   2.285088     0.91   0.364    -2.404559    6.552822
previousPM~1 |   .3449009   .3383798     1.02   0.308    -.3183114    1.008113
previous_c~1 |   .2361982   .1454148     1.62   0.104    -.0488096     .521206
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

. *     ****************************************************************  *;
. *       What if we had used all governments in Western Europe?          *;
. *     ****************************************************************  *;
. use "C:\fsu2010\formateur\ajps\submission2\replication\analysis\europe_all.dta", clear;

. *     ****************************************************************  *;
. *       Model 1: Western Europe                                         *;
. *     ****************************************************************  *;
. mixlogit choice largestparty_east party_seatshare_east,  
>     rand (largestparty party_seatshare) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -388.95988  (not concave)
Iteration 1:   log likelihood = -359.43993  
Iteration 2:   log likelihood = -352.95587  (not concave)
Iteration 3:   log likelihood = -352.27833  
Iteration 4:   log likelihood = -351.06843  
Iteration 5:   log likelihood = -350.97716  
Iteration 6:   log likelihood = -350.97247  
Iteration 7:   log likelihood = -350.97244  

Mixed logit model                                 Number of obs   =       2856
                                                  LR chi2(2)      =      10.73
Log likelihood = -350.97244                       Prob > chi2     =     0.0047

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~ast |  -.5315489   .7740904    -0.69   0.492    -2.048738    .9856403
party_se~ast |   .1115689   .0570306     1.96   0.050    -.0002091    .2233469
largestparty |   .9695774   .5116214     1.90   0.058     -.033182    1.972337
party_seat~e |   .1461963   .0187961     7.78   0.000     .1093566    .1830361
-------------+----------------------------------------------------------------
SD           |
largestparty |  -2.976469   .9799995    -3.04   0.002    -4.897233   -1.055705
party_seat~e |  -.0059255   .0629528    -0.09   0.925    -.1293107    .1174598
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.  *     ****************************************************************  *;
. *       Model 2: Western Europe                                         *;
. *     ****************************************************************  *;
. mixlogit choice largestparty_east party_seatshare_east medianparty1_east,  
>     rand (largestparty party_seatshare medianparty1) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -340.91127  (not concave)
Iteration 1:   log likelihood = -314.92774  
Iteration 2:   log likelihood = -309.85493  
Iteration 3:   log likelihood = -309.71543  
Iteration 4:   log likelihood = -309.06989  
Iteration 5:   log likelihood = -309.02172  
Iteration 6:   log likelihood = -309.01529  
Iteration 7:   log likelihood = -309.01485  
Iteration 8:   log likelihood = -309.01484  

Mixed logit model                                 Number of obs   =       2648
                                                  LR chi2(3)      =       5.06
Log likelihood = -309.01484                       Prob > chi2     =     0.1675

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~ast |  -1.344145   .8486993    -1.58   0.113    -3.007565    .3192751
party_se~ast |   .2259785   .0840709     2.69   0.007     .0612027    .3907544
medianpa~ast |  -1.408323      .7128    -1.98   0.048    -2.805386    -.011261
largestparty |   .5791821   .3184735     1.82   0.069    -.0450145    1.203379
party_seat~e |   .1216239   .0158004     7.70   0.000     .0906558     .152592
medianparty1 |   1.044235   .2120995     4.92   0.000     .6285272    1.459942
-------------+----------------------------------------------------------------
SD           |
largestparty |   -1.88929   .6336367    -2.98   0.003    -3.131195   -.6473849
party_seat~e |   .0017585   .0421847     0.04   0.967    -.0809221     .084439
medianparty1 |   .1622575   3.046849     0.05   0.958    -5.809457    6.133972
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.      *     ****************************************************************  *;
. *       Model 3: Western Europe                                         *;
. *     ****************************************************************  *;
. mixlogit choice largestparty_east party_seatshare_east medianparty1_east presidentparty_east,  
>     rand (largestparty party_seatshare medianparty1 presidentparty) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood =   -340.546  (not concave)
Iteration 1:   log likelihood = -314.20872  (not concave)
Iteration 2:   log likelihood = -313.05649  (not concave)
Iteration 3:   log likelihood = -308.02324  (not concave)
Iteration 4:   log likelihood = -307.95158  
Iteration 5:   log likelihood = -306.27216  
Iteration 6:   log likelihood = -306.07638  
Iteration 7:   log likelihood = -306.07459  
Iteration 8:   log likelihood = -306.07459  

Mixed logit model                                 Number of obs   =       2648
                                                  LR chi2(4)      =       9.20
Log likelihood = -306.07459                       Prob > chi2     =     0.0563

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~ast |  -1.572961   .9901531    -1.59   0.112    -3.513625    .3677033
party_se~ast |    .252756   .0930938     2.72   0.007     .0702954    .4352166
medianpa~ast |  -1.316432   .7919021    -1.66   0.096    -2.868532    .2356673
presi~y_east |  -.5386789   1.074745    -0.50   0.616     -2.64514    1.567782
largestparty |   .7619347   .4057971     1.88   0.060    -.0334129    1.557282
party_seat~e |    .127978   .0178806     7.16   0.000     .0929326    .1630233
medianparty1 |   1.093655   .2386635     4.58   0.000     .6258835    1.561427
presidentp~y |    .732783   .4302486     1.70   0.089    -.1104887    1.576055
-------------+----------------------------------------------------------------
SD           |
largestparty |   2.538181   .7974035     3.18   0.001     .9752993    4.101064
party_seat~e |   .0035974   .0386314     0.09   0.926    -.0721186    .0793135
medianparty1 |   .3360877   .9698535     0.35   0.729     -1.56479    2.236966
presidentp~y |   1.221739   .8473777     1.44   0.149    -.4390905    2.882569
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.           *     ****************************************************************  *;
. *       Model 4: Western Europe                                         *;
. *     ****************************************************************  *;
. mixlogit choice largestparty_east party_seatshare_east medianparty1_east presidentparty_east previousPM_east,  
>     rand (largestparty party_seatshare medianparty1 presidentparty previousPM) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -330.26997  (not concave)
Iteration 1:   log likelihood = -325.25795  (not concave)
Iteration 2:   log likelihood = -298.63674  (not concave)
Iteration 3:   log likelihood = -298.11525  
Iteration 4:   log likelihood = -296.57287  
Iteration 5:   log likelihood = -292.32305  (not concave)
Iteration 6:   log likelihood = -291.94059  
Iteration 7:   log likelihood = -291.63694  
Iteration 8:   log likelihood = -291.63175  
Iteration 9:   log likelihood = -291.63174  

Mixed logit model                                 Number of obs   =       2648
                                                  LR chi2(5)      =      16.53
Log likelihood = -291.63174                       Prob > chi2     =     0.0055

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~ast |  -1.552732   1.075501    -1.44   0.149    -3.660675      .55521
party_se~ast |   .2414079   .0991627     2.43   0.015     .0470526    .4357632
medianpa~ast |  -1.795314   .9082835    -1.98   0.048    -3.575516   -.0151106
presi~y_east |   .0307693   1.152353     0.03   0.979    -2.227801     2.28934
previ~M_east |   .2766624   .8782082     0.32   0.753    -1.444594    1.997919
largestparty |   .7928743   .4525074     1.75   0.080    -.0940239    1.679773
party_seat~e |    .119923    .020287     5.91   0.000     .0801612    .1596849
medianparty1 |   1.165494   .2914726     4.00   0.000      .594218    1.736769
presidentp~y |    .746532   .4903033     1.52   0.128    -.2144447    1.707509
  previousPM |   1.143188    .395445     2.89   0.004     .3681297    1.918246
-------------+----------------------------------------------------------------
SD           |
largestparty |   2.085399   1.223342     1.70   0.088    -.3123077    4.483106
party_seat~e |  -.0096495   .0298351    -0.32   0.746    -.0681252    .0488262
medianparty1 |   .0526748   .9808058     0.05   0.957    -1.869669    1.975019
presidentp~y |   .6972599   1.367347     0.51   0.610     -1.98269     3.37721
  previousPM |  -2.610812   .8608393    -3.03   0.002    -4.298026   -.9235982
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.     *     ****************************************************************  *;
. *       Model 5: Western Europe                                         *;
. *     ****************************************************************  *;
. mixlogit choice largestparty_east party_seatshare_east medianparty1_east presidentparty_east previousPM_east,   
>     rand (largestparty party_seatshare medianparty1 presidentparty previousPM president_investiture president_direct) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -326.45461  (not concave)
Iteration 1:   log likelihood = -302.83835  
Iteration 2:   log likelihood = -290.00637  (not concave)
Iteration 3:   log likelihood = -289.93731  
Iteration 4:   log likelihood = -288.71189  
Iteration 5:   log likelihood = -288.63433  
Iteration 6:   log likelihood = -288.63282  
Iteration 7:   log likelihood = -288.63282  

Mixed logit model                                 Number of obs   =       2648
                                                  LR chi2(7)      =      14.75
Log likelihood = -288.63282                       Prob > chi2     =     0.0393

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~ast |  -1.559722   1.029528    -1.51   0.130     -3.57756     .458115
party_se~ast |   .2455531   .0983096     2.50   0.012     .0528698    .4382363
medianpa~ast |  -1.707935   .8946852    -1.91   0.056    -3.461486    .0456155
presi~y_east |   .8228761   1.233621     0.67   0.505    -1.594978     3.24073
previ~M_east |   .4412214   .8314856     0.53   0.596     -1.18846    2.070903
largestparty |   .7665859   .4163935     1.84   0.066    -.0495304    1.582702
party_seat~e |   .1175887   .0195858     6.00   0.000     .0792012    .1559762
medianparty1 |   1.156793   .2838728     4.08   0.000     .6004125    1.713173
presidentp~y |   3.382548   1.355531     2.50   0.013     .7257554    6.039341
  previousPM |   1.059679   .3695752     2.87   0.004     .3353245    1.784033
president_~e |  -2.237243   1.144517    -1.95   0.051    -4.480455    .0059696
president~ct |  -2.417167   1.176757    -2.05   0.040    -4.723568   -.1107664
-------------+----------------------------------------------------------------
SD           |
largestparty |   1.854568   1.147285     1.62   0.106    -.3940688    4.103205
party_seat~e |  -.0061321   .0308063    -0.20   0.842    -.0665114    .0542472
medianparty1 |   .2446315   .8943942     0.27   0.784    -1.508349    1.997612
presidentp~y |   .0684241   1.345339     0.05   0.959    -2.568391    2.705239
  previousPM |  -2.412475   .7977079    -3.02   0.002    -3.975953   -.8489958
president_~e |   -.035648   1.928481    -0.02   0.985      -3.8154    3.744104
president~ct |   .5971138   1.144291     0.52   0.602    -1.645655    2.839883
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

. mixlogit choice largestparty_east party_seatshare_east medianparty1_east presidentparty_east previousPM_east,    
>     rand (largestparty party_seatshare medianparty1 presidentparty previousPM president_investiture president_direct) ln(0) nrep(125) group(cabinetcode) 
>     from(largestparty_east =-1.403309 party_seatshare_east =0.2506177 medianparty1_east =-1.441322 presidentparty_east =0.9270295 
>     previousPM_east =-1.228084 largestparty=1.002409 party_seatshare=0.1165616 medianparty1=1.081568 presidentparty=3.408346 
>     previousPM=0.7209862 president_investiture= -2.267492 president_direct= -2.237836);

Iteration 0:   log likelihood = -332.02323  (not concave)
Iteration 1:   log likelihood = -299.32858  (not concave)
Iteration 2:   log likelihood = -298.19095  (not concave)
Iteration 3:   log likelihood = -296.38156  
Iteration 4:   log likelihood = -289.51354  (not concave)
Iteration 5:   log likelihood = -289.11097  
Iteration 6:   log likelihood = -288.58444  
Iteration 7:   log likelihood = -288.57007  
Iteration 8:   log likelihood = -288.56973  
Iteration 9:   log likelihood = -288.56973  

Mixed logit model                                 Number of obs   =       2648
                                                  LR chi2(7)      =      14.88
Log likelihood = -288.56973                       Prob > chi2     =     0.0376

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~ast |  -1.512833   1.030244    -1.47   0.142    -3.532075     .506409
party_se~ast |   .2447571   .0986838     2.48   0.013     .0513403    .4381738
medianpa~ast |  -1.735683   .8962724    -1.94   0.053    -3.492344    .0209789
presi~y_east |   .8211148   1.213622     0.68   0.499     -1.55754    3.199769
previ~M_east |   .5002856    .841302     0.59   0.552    -1.148636    2.149207
largestparty |    .751493   .4187948     1.79   0.073    -.0693298    1.572316
party_seat~e |   .1170547   .0190827     6.13   0.000     .0796534    .1544561
medianparty1 |   1.160699   .2836657     4.09   0.000     .6047246    1.716674
presidentp~y |   3.401885   1.346396     2.53   0.012     .7629971    6.040773
  previousPM |    1.03393   .3603442     2.87   0.004      .327668    1.740191
president_~e |  -2.263431   1.126584    -2.01   0.045    -4.471495   -.0553674
president~ct |  -2.408913   1.176717    -2.05   0.041    -4.715237   -.1025896
-------------+----------------------------------------------------------------
SD           |
largestparty |    1.85794   1.153009     1.61   0.107    -.4019163    4.117796
party_seat~e |   .0040203   .0303615     0.13   0.895     -.055487    .0635277
medianparty1 |  -.0013724   1.092375    -0.00   0.999    -2.142387    2.139642
presidentp~y |  -.1010617   2.805072    -0.04   0.971    -5.598902    5.396779
  previousPM |   2.399985   .7705269     3.11   0.002     .8897797     3.91019
president_~e |  -.1982515   1.296824    -0.15   0.878    -2.739979    2.343476
president~ct |  -.2966179   1.658189    -0.18   0.858    -3.546608    2.953372
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.        *     ****************************************************************  *;
. *       Model 6: Western Europe                                         *;
. *     ****************************************************************  *;
. mixlogit choice largestparty_east party_seatshare_east medianparty1_east presidentparty_east previousPM_east previousPM_conflict_east previousPM_gain1_east previous_cabinet_gain1_east,  
>     rand (largestparty party_seatshare medianparty1 presidentparty previousPM president_investiture president_direct previousPM_conflict previousPM_gain1 previous_cabinet_gain1) 
>     ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -279.64736  (not concave)
Iteration 1:   log likelihood = -279.35948  (not concave)
Iteration 2:   log likelihood = -252.10719  (not concave)
Iteration 3:   log likelihood = -250.80742  
Iteration 4:   log likelihood = -243.62181  (not concave)
Iteration 5:   log likelihood = -242.86598  
Iteration 6:   log likelihood = -241.70201  (not concave)
Iteration 7:   log likelihood = -241.63262  
Iteration 8:   log likelihood = -241.62452  
Iteration 9:   log likelihood = -241.59713  
Iteration 10:  log likelihood = -241.58576  
Iteration 11:  log likelihood = -241.58557  
Iteration 12:  log likelihood = -241.58557  

Mixed logit model                                 Number of obs   =       2405
                                                  LR chi2(10)     =      25.15
Log likelihood = -241.58557                       Prob > chi2     =     0.0051

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestp~ast |  -1.830528    1.63005    -1.12   0.261    -5.025368    1.364313
party_se~ast |   .2841378   .1358168     2.09   0.036     .0179418    .5503337
medianpa~ast |  -2.966238   1.509352    -1.97   0.049    -5.924514   -.0079612
presi~y_east |  -.6539151   1.742592    -0.38   0.707    -4.069332    2.761502
previ~M_east |    .239587    1.97257     0.12   0.903     -3.62658    4.105753
previ~t_east |   .2252191   2.536536     0.09   0.929      -4.7463    5.196739
previousPM.. |  -.1533073   .4347585    -0.35   0.724    -1.005418    .6988038
previous_c.. |   .0039077   .1807898     0.02   0.983    -.3504337    .3582491
largestparty |   1.550876   .8700085     1.78   0.075    -.1543091    3.256061
party_seat~e |   .1382921   .0287253     4.81   0.000     .0819916    .1945926
medianparty1 |   1.457735   .3784479     3.85   0.000     .7159911    2.199479
presidentp~y |   5.211777   2.212656     2.36   0.019     .8750503    9.548504
  previousPM |   3.218465   1.154851     2.79   0.005     .9549983    5.481931
president_~e |  -2.824216   1.940446    -1.46   0.146     -6.62742    .9789886
president~ct |  -3.960781   1.890669    -2.09   0.036    -7.666424   -.2551375
previousP~ct |  -1.881149   1.146848    -1.64   0.101     -4.12893    .3666318
previousPM~1 |   .6471647   .2468476     2.62   0.009     .1633523    1.130977
previous_c~1 |  -.0039805   .0488304    -0.08   0.935    -.0996864    .0917253
-------------+----------------------------------------------------------------
SD           |
largestparty |  -3.791861   1.474613    -2.57   0.010    -6.682049   -.9016726
party_seat~e |   .0240049   .0380337     0.63   0.528    -.0505398    .0985496
medianparty1 |   .1451377   .7016841     0.21   0.836    -1.230138    1.520413
presidentp~y |  -.1511695   1.068883    -0.14   0.888    -2.246141    1.943802
  previousPM |    2.01053   1.504228     1.34   0.181    -.9377028    4.958764
president_~e |  -.2370884   1.422395    -0.17   0.868    -3.024931    2.550754
president~ct |  -.1928782   1.455089    -0.13   0.895    -3.044801    2.659044
previousP~ct |  -1.304916   2.547699    -0.51   0.609    -6.298315    3.688482
previousPM~1 |   .2249948   .3680095     0.61   0.541    -.4962907    .9462802
previous_c~1 |   .2799569   .1825515     1.53   0.125    -.0778375    .6377512
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.     clear;

. *     ****************************************************************  *;
. *       What if we had used the CMP data to identify the median party   *;
. *       in Western Europe? Let's just look at the West Europe data.     *;
. *     ****************************************************************  *;
. use "C:\fsu2010\formateur\ajps\submission2\replication\analysis\westeurope2_CMP.dta", clear;

.   *     ****************************************************************  *;
. *       Model 2: Western Europe (CMP version)                           *;
. *     ****************************************************************  *;
. mixlogit choice, rand (largestparty party_seatshare medianparty_CMP) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -279.37176  (not concave)
Iteration 1:   log likelihood = -252.69779  (not concave)
Iteration 2:   log likelihood = -251.98353  
Iteration 3:   log likelihood = -249.74762  
Iteration 4:   log likelihood = -249.29596  
Iteration 5:   log likelihood = -249.27787  
Iteration 6:   log likelihood = -249.27781  
Iteration 7:   log likelihood = -249.27781  

Mixed logit model                                 Number of obs   =       1929
                                                  LR chi2(3)      =       7.79
Log likelihood = -249.27781                       Prob > chi2     =     0.0505

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestparty |   1.012599   .5952943     1.70   0.089    -.1541567    2.179354
party_seat~e |   .1384565   .0265855     5.21   0.000     .0863499     .190563
medianpart~P |   .3047731   .2788555     1.09   0.274    -.2417737    .8513199
-------------+----------------------------------------------------------------
SD           |
largestparty |  -2.966494   1.233963    -2.40   0.016    -5.385017   -.5479704
party_seat~e |  -.0179372   .0481956    -0.37   0.710    -.1123987    .0765244
medianpart~P |   .0797904   .5731243     0.14   0.889    -1.043513    1.203093
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.      *     ****************************************************************  *;
. *       Model 3: Western Europe (CMP version)                           *;
. *     ****************************************************************  *;
. mixlogit choice, rand (largestparty party_seatshare medianparty_CMP presidentparty) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -279.15877  (not concave)
Iteration 1:   log likelihood = -252.04778  (not concave)
Iteration 2:   log likelihood = -251.96265  (not concave)
Iteration 3:   log likelihood = -250.93304  
Iteration 4:   log likelihood = -246.87147  
Iteration 5:   log likelihood = -245.63285  
Iteration 6:   log likelihood = -245.21285  
Iteration 7:   log likelihood = -245.19958  
Iteration 8:   log likelihood = -245.19932  
Iteration 9:   log likelihood = -245.19932  

Mixed logit model                                 Number of obs   =       1929
                                                  LR chi2(4)      =      14.54
Log likelihood = -245.19932                       Prob > chi2     =     0.0058

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestparty |   1.836012   1.157728     1.59   0.113    -.4330936    4.105117
party_seat~e |   .1521019    .027508     5.53   0.000     .0981873    .2060166
medianpart~P |   .2819878   .3521512     0.80   0.423     -.408216    .9721916
presidentp~y |   1.101141   .6899127     1.60   0.110    -.2510633    2.453345
-------------+----------------------------------------------------------------
SD           |
largestparty |  -4.892511   2.179932    -2.24   0.025    -9.165098    -.619923
party_seat~e |  -.0142372   .0611903    -0.23   0.816     -.134168    .1056936
medianpart~P |   .0942723   .7066173     0.13   0.894    -1.290672    1.479217
presidentp~y |   1.670622   1.275159     1.31   0.190    -.8286436    4.169888
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.       *     ****************************************************************  *;
. *       Model 4: Western Europe (CMP version)                           *;
. *     ****************************************************************  *;
. mixlogit choice, rand (largestparty party_seatshare medianparty_CMP presidentparty previousPM) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -273.59927  (not concave)
Iteration 1:   log likelihood = -246.44633  (not concave)
Iteration 2:   log likelihood = -245.32662  (not concave)
Iteration 3:   log likelihood = -244.41889  
Iteration 4:   log likelihood = -239.72996  
Iteration 5:   log likelihood =  -239.4091  
Iteration 6:   log likelihood = -239.12823  
Iteration 7:   log likelihood = -239.10144  
Iteration 8:   log likelihood =  -239.0983  
Iteration 9:   log likelihood =  -239.0983  

Mixed logit model                                 Number of obs   =       1929
                                                  LR chi2(5)      =      15.98
Log likelihood =  -239.0983                       Prob > chi2     =     0.0069

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestparty |   1.214094   .8024285     1.51   0.130    -.3586364    2.786825
party_seat~e |   .1370394    .023248     5.89   0.000     .0914742    .1826046
medianpart~P |   .3278882   .3519739     0.93   0.352    -.3619679    1.017744
presidentp~y |   .8413525   .6726433     1.25   0.211    -.4770042    2.159709
  previousPM |     1.0686   .5174309     2.07   0.039     .0544546    2.082746
-------------+----------------------------------------------------------------
SD           |
largestparty |   -3.62166   2.022427    -1.79   0.073    -7.585545    .3422241
party_seat~e |   .0059674   .0461298     0.13   0.897    -.0844453    .0963801
medianpart~P |   .1281132   .6829974     0.19   0.851    -1.210537    1.466764
presidentp~y |   1.593967   1.758383     0.91   0.365      -1.8524    5.040334
  previousPM |  -2.032331   1.571889    -1.29   0.196    -5.113177    1.048515
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.  *     ****************************************************************  *;
. *       Model 5: Western Europe (CMP version)                      *;
. *     ****************************************************************  *;
. mixlogit choice, rand (largestparty party_seatshare medianparty_CMP presidentparty previousPM president_investiture president_direct) ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -269.82755  (not concave)
Iteration 1:   log likelihood = -243.45051  (not concave)
Iteration 2:   log likelihood = -243.43121  (not concave)
Iteration 3:   log likelihood = -241.20339  
Iteration 4:   log likelihood = -237.57078  
Iteration 5:   log likelihood = -237.02546  
Iteration 6:   log likelihood = -236.38541  
Iteration 7:   log likelihood =  -235.9861  
Iteration 8:   log likelihood = -235.95232  
Iteration 9:   log likelihood = -235.95096  
Iteration 10:  log likelihood = -235.95096  

Mixed logit model                                 Number of obs   =       1929
                                                  LR chi2(7)      =      15.19
Log likelihood = -235.95096                       Prob > chi2     =     0.0337

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestparty |   1.335153   .7569007     1.76   0.078    -.1483449    2.818651
party_seat~e |   .1382015   .0223379     6.19   0.000       .09442    .1819829
medianpart~P |    .322714   .3471777     0.93   0.353    -.3577418     1.00317
presidentp~y |   5.154281   2.074282     2.48   0.013     1.088763    9.219798
  previousPM |   1.060191   .4751292     2.23   0.026     .1289554    1.991428
president_~e |  -3.008294   1.569572    -1.92   0.055    -6.084599    .0680102
president_~t |  -3.864381   1.852635    -2.09   0.037    -7.495478   -.2332834
-------------+----------------------------------------------------------------
SD           |
largestparty |   3.819252   1.551407     2.46   0.014     .7785493    6.859954
party_seat~e |  -.0024369   .0394076    -0.06   0.951    -.0796744    .0748005
medianpart~P |   .0193476   .7683954     0.03   0.980     -1.48668    1.525375
presidentp~y |   .1627113   1.387894     0.12   0.907    -2.557512    2.882934
  previousPM |   2.113412   1.186358     1.78   0.075    -.2118061     4.43863
president_~e |  -.2758251   2.644793    -0.10   0.917    -5.459525    4.907874
president_~t |   .4595521    1.62181     0.28   0.777    -2.719136    3.638241
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

.      *     ****************************************************************  *;
. *       Model 6: Western Europe (CMP version)                           *;
. *     ****************************************************************  *;
. mixlogit choice, rand (largestparty party_seatshare medianparty_CMP presidentparty previousPM president_investiture president_direct previousPM_conflict previousPM_gain1 previous_cabinet_gain1) 
>     ln(0) nrep(125) group(cabinetcode);

Iteration 0:   log likelihood = -232.94793  (not concave)
Iteration 1:   log likelihood = -209.42098  (not concave)
Iteration 2:   log likelihood = -208.87633  
Iteration 3:   log likelihood = -206.17853  (not concave)
Iteration 4:   log likelihood = -205.55968  
Iteration 5:   log likelihood = -205.13317  (not concave)
Iteration 6:   log likelihood = -204.60685  (not concave)
Iteration 7:   log likelihood = -203.55242  
Iteration 8:   log likelihood = -200.31641  (not concave)
Iteration 9:   log likelihood = -200.27774  (not concave)
Iteration 10:  log likelihood = -200.22643  (not concave)
Iteration 11:  log likelihood = -199.79241  (not concave)
Iteration 12:  log likelihood = -199.33114  (not concave)
Iteration 13:  log likelihood = -199.01246  
Iteration 14:  log likelihood = -198.26421  
Iteration 15:  log likelihood = -198.12395  (not concave)
Iteration 16:  log likelihood = -198.11901  
Iteration 17:  log likelihood = -198.04153  
Iteration 18:  log likelihood = -198.03828  
Iteration 19:  log likelihood = -198.03825  
Iteration 20:  log likelihood = -198.03824  

Mixed logit model                                 Number of obs   =       1789
                                                  LR chi2(10)     =      24.61
Log likelihood = -198.03824                       Prob > chi2     =     0.0061

------------------------------------------------------------------------------
      choice |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Mean         |
largestparty |   3.583378   2.110273     1.70   0.089    -.5526806    7.719437
party_seat~e |   .1588514   .0285639     5.56   0.000     .1028671    .2148356
medianpart~P |   .1397239   .5007454     0.28   0.780     -.841719    1.121167
presidentp~y |    6.69435   2.961572     2.26   0.024     .8897755    12.49892
  previousPM |   4.307266   1.832577     2.35   0.019       .71548    7.899051
president_~e |  -3.256324   2.613265    -1.25   0.213    -8.378229    1.865582
president_~t |  -5.486502    2.77478    -1.98   0.048    -10.92497   -.0480329
previousPM~t |  -3.543503   2.038824    -1.74   0.082    -7.539525    .4525193
previousPM~1 |   .8770208   .4230308     2.07   0.038     .0478956    1.706146
previous_c~1 |   .0303707   .0653379     0.46   0.642    -.0976892    .1584305
-------------+----------------------------------------------------------------
SD           |
largestparty |   6.624418   2.892172     2.29   0.022     .9558648    12.29297
party_seat~e |  -.0026719   .0670111    -0.04   0.968    -.1340113    .1286675
medianpart~P |  -.1695708   .8991917    -0.19   0.850    -1.931954    1.592813
presidentp~y |  -.2349553   1.362096    -0.17   0.863    -2.904614    2.434704
  previousPM |   2.210532   2.340913     0.94   0.345    -2.377572    6.798636
president_~e |  -.1467552   1.705003    -0.09   0.931    -3.488499    3.194989
president_~t |   .0103596   2.591423     0.00   0.997    -5.068736    5.089455
previousPM~t |   .3259604   8.847122     0.04   0.971    -17.01408      17.666
previousPM~1 |   .2857144    .270445     1.06   0.291    -.2443482    .8157769
previous_c~1 |   .4719874   .2422222     1.95   0.051    -.0027594    .9467341
------------------------------------------------------------------------------
The sign of the estimated standard deviations is irrelevant: interpret them as
being positive

. *     ****************************************************************  *;
. *       No evidence that median party has a higher probability of       *;
. *       becoming the PM party in W. Europe once we use CMP to identify  *;
. *       the median party.                                               *;
. *     ****************************************************************  *;
. log close;
      name:  <unnamed>
       log:  C:\fsu2010\formateur\ajps\submission2\replication\analysis\stata_mxl_robustness.log
  log type:  text
 closed on:  21 Mar 2011, 14:00:05
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
